Good afternoon, welcome to TPI Composites' third quarter 2018 earnings conference call. Today's call is being recorded, we have allocated one hour for prepared remarks and questions and answers. At this time, I'd like to turn the conference over to Anthony Rasmus, investor relations for TPI Composites. Thank you. You may begin.
Thank you, operator. I'd like to welcome everyone to TPI Composites' third quarter 2018 earnings call. We will be making forward-looking statements during this call based on current expectations and assumptions, which are subject to risk and uncertainties. Actual results could differ materially from our forward-looking statements if any of our key assumptions are incorrect or because of other factors discussed in today's earnings news release and the comments made during this conference call, or in our latest reports and filings with the Securities and Exchange Commission, each of which can be found on our website, www.tpicomposites.com. We do not undertake any duty to update any forward-looking statements. Today's presentation also includes references to non-GAAP financial measures. You should refer to the information contained in the slides accompanying today's presentation for definitional information and reconciliations of historical non-GAAP measures to the closest GAAP financial measure.
With that, let me turn the call over to Steven Lockard, TPI Composites' President and Chief Executive Officer.
Good afternoon, everyone, thank you for joining our third quarter 2018 earnings call. I'm joined today by William Siwek, our Chief Financial Officer. I'll begin with some highlights from the quarter, followed by a brief update of the wind market. I'll then turn the call over to Bill to review our financial results, our full year 2018 outlook, introduce 2019 guidance, and provide preliminary 2020 financial targets before we open up the call for Q&A. Please turn to slide five. In Q3, we continued the investment we discussed in the second quarter in new startups, as well as transitioning lines to new, larger blades. We believe this investment will provide a foundation for our continued growth in 2019 and beyond.
We are tracking to plan, notwithstanding the significant amount of effort and complexity involved in the startup and transition process, and our results are on pace to be in line with our 2018 guidance. Our net sales for Q3 were $255 million, and adjusted EBITDA was $17.6 million. Due to large number of transitions in process, which reduced our blade volumes, our sales were relatively flat. The impact of both startups and transitions resulted in our adjusted EBITDA declining quarter-over-quarter, specifically due to the lost contribution margin dollars from lower blade volume during the transitions, coupled with high cost of startups. However, gross margins and adjusted EBITDA margins before the impact of startup and transition costs both remained north of 15% as a result of continued reductions in manufacturing cycle times, improvements in productivity, material cost-out efforts and the strength of the US dollar.
Our customers continue to invest with TPI in adding new outsourced blade capacity ahead of the new line guidance we provided for 2018. In addition, they're tooling up new larger blade models more quickly than initially planned in order to aggressively drive down LCOE in response to economically driven global auction and tender processes. On the transportation side of our business, today we are announcing an investment of approximately $11.5 million in 2019 to develop a highly automated pilot manufacturing line for the electric vehicle market. This pilot line will be located adjacent to our recently opened plant in Newton, Iowa, where we began manufacturing bus bodies for Proterra at the end of the second quarter. This investment will enable us to further develop our technology, create defensible product and process IP, and demonstrate our capability to manufacture composite components cost effectively at automotive volume rates.
This pilot line will also help our current and potential customers to de-risk the decision-making process to commit to TPI for high-volume manufacturing programs in the future. Turning back to our wind blade business, since the beginning of 2018, we have added a net of nine new lines to bring our total dedicated lines under long-term contracts as of today to 51. The net of these transactions, as well as a few amendments to existing supply agreements, represent potential contract revenue of up to $2.6 billion over the terms of these agreements. Furthermore, we have now closed on 12 new lines so far in 2018, 10 of which were in our prioritized pipeline at the start of the year, putting us at the midpoint of our guidance range for new lines in 2018 of between 10 and 14.
We now have a total potential contract value of up to $6.3 billion extending through 2023. We continue to develop our robust wind pipeline of global opportunities with current and new customers in both onshore and offshore blades. At the beginning of the year, our prioritized pipeline was 24 lines, 10 of which we have closed on so far this year. We have now updated our pipeline to include lines we intend to close by the end of 2020. This updated prioritized pipeline now includes a total of 23 lines, or an incremental nine lines to be converted between today and the end of 2020. We are confident in our ability to convert this pipeline and continue to be in active negotiations with the expectation of closing additional lines prior to the end of 2018. Please turn to slide six.
As we've talked about previously, 2018 is an investment year for TPI, and we have held our estimate of lines in startup during the year at 17 and updated the lines in transition during the year to 15, down two lines from Q2 as those moved into 2019 to accommodate a timing change by one of our customers. We discussed the impact of the increased volume of startups and transitions on near-term profitability last quarter, and our adjusted EBITDA guidance for 2018 remains unchanged. What we may have not clearly articulated is the incremental future potential revenue as a result of these costs and the anticipated corresponding earnings that follow that revenue. As we think about the startup costs that we expect to incur this year and in 2019 for lines under contract today, the potential revenue opportunity under those contracts is approximately $3.8 billion from 2019 through 2023.
With respect to the transition costs that we expect to incur this year and in 2019 for lines under contract today, the incremental potential revenue opportunity under those contracts is approximately $500 million from 2019 through 2022. Given our historical returns on invested capital from startups and our rigorous investment policy, startup costs translate to growth and future potential profitability, and we believe the same is true with most transitions. Notwithstanding the fact that done the right way, transitions help to fuel future growth, not only for TPI, but for the wind industry as a whole. TPI and our customers would prefer the transition slow to a more normalized pace. We have and will continue to evaluate every transition request from our customers to ensure that it's in the best interest of TPI, our stakeholders, and of course, our customers.
While we've had some execution challenges and delays relating to both startups and transitions this year, we're getting better with our customer support at both startups and transitions, making them happen faster and therefore less costly. In the near future, we anticipate more families of blade models designed for rapid changes. For example, tip changes instead of full mold changes, as well as modular or split blades. We also expect blade transitions to reduce over time as LCOE drops below competitive technologies, blade lengths reach practical limits for ground transportation, and the incremental costs of new blade models outweigh the incremental LCOE impact. Our growth strategy remains intact, and we continue to see traction as we diversify our sources of revenue across customers, geographies, and non-wind markets.
We plan to execute on this strategy and take advantage of the growth in the global wind market, stability in the U.S. wind market, and the ongoing wind blade outsourcing trend. Turning to slide seven and eight. As of today, our long-term supply agreements provide potential revenue of up to $6.3 billion through 2023. At this time last year, our potential revenue under our supply agreements was approximately $4.4 billion. We have increased that amount by approximately $1.9 billion net of the impact of approximately $945 million of billings since that time. In other words, contract value added since last year at this time through new deals, amendments, and blade transitions prior to considering what we've realized in total billings over the last year is over $3 billion.
The minimum guaranteed volume under our supply agreements has grown to approximately $4.3 billion, up from $2.7 billion at this time last year. Turning to the wind market. Global annual wind power capacity additions are expected to average nearly 67 GW between 2017 and 2027, or a 10-year CAGR of 8.2%, according to MAKE Consulting. This forecast also estimates that the top 20 emerging markets will grow at a CAGR of 26.5% between 2017 and 2027. Our strategy is to continue to diversify our manufacturing footprint to take advantage of growth in both emerging and mature markets and leverage our low-cost hubs to not be too dependent on any one market. We believe we remain well-positioned to execute this strategy and serve global demand from our facilities in the U.S., China, Mexico and Turkey.
We also expect to enter a new geography during 2019 to accommodate customer demand and to diversify our global footprint and supply chain even further. We expect this global growth to continue to drive the outsourcing trend we've seen over the last 10 years. The U.S. market outlook over the next several years is strengthening, with expected annual installations averaging 11.1 GW through 2021, and then averaging just over eight GW from 2022 through 2025, according to UBS. We, like many participants in the wind and utilities industries, believe that the economics of wind, along with demand from both retail and industrial customers, the electrification of the vehicle fleet, and decarbonization initiatives by utilities, will continue to drive wind penetration long after the current PTC sunsets in 2023.
Before I turn the call over to Will, I'll touch briefly on the recent tariffs levied by the U.S. government on goods, including wind blades and other turbine components imported from China. Wind blades, as well as other components used in wind turbines, are some of the many products included in the list of those to be covered initially by a 10% tariff, which then increases to 25% on January 1, 2019. A number of turbine OEMs import blades and components to the U.S. from China when the U.S. market demand is strong, including blades manufactured by TPI. Although we don't know how long these tariffs may be in place, initial estimates by others suggest the overall impact on the levelized cost of wind energy in the U.S. could be 5%-10%, including the tariffs on steel.
Today, less than 15% of the blades we produce globally are imported by our customers into the U.S. from China, we recognize that the added cost of the tariff may result in our customers shifting which TPI or other factory they source the U.S. blades from. The benefit of our global manufacturing footprint is that it can allow our customers to shift volumes to best meet their cost and delivery requirements. We still have strong demand on our China facilities in 2019, even though many of the blades will be shipped to other locations. This is further demonstration why having world-class manufacturing hubs to serve large geographies cost effectively is important for our customers and provides us with an advantage over most other blade manufacturers. With that, let me turn the call over to Bill.
Thanks, Steve. Please refer to slides 10 and 11. Net sales for the quarter were $255 million, or an increase of 0.6% compared to the same period in 2017. Net sales of wind blades were $234.9 million for the quarter as compared to $238.1 million in the same period of 2017. The decrease was primarily driven by a 19.1% decrease in the number of wind blades produced during the third quarter of 2018 compared to the third quarter of 2017 as a result of the increase in lines in transition, the loss to volume from a contract that expired at the end of 2017, and a delayed customer startup, which was partially offset by a higher average sales price due to the mix of wind blade models produced during the quarter compared to the same period in 2017, an increase in non-blade revenue and by foreign currency fluctuations.
Total billings for the third quarter decreased by $15.7 million, or 6.1%, to $240.7 million compared to the same period in 2017. The impact of the fluctuating US dollar against the euro in our Turkey operations and the Chinese RMB in our China operations on consolidated net sales and total billings for the three months ended September 30, 2018, was a net decrease of 1.1% and 1.2% respectively as compared to the same period in 2017. Gross profit for the quarter totaled $17 million, a decrease of $13.3 million over the same period of 2017, and our gross profit margin decreased to 6.7%.
The lower gross margin was primarily driven by the loss of contribution margin dollars from lower blade volume during transitions, an increase in startup costs of $6.6 million, and an increase in transition costs of $2.4 million compared to the same period a year ago, offset by favorable foreign currency movements. On a constant currency basis and before startup and transition costs, gross margin was 11.3% compared to 16.8% in Q3 of 2017. This decline was primarily due to the impact of an increased amount of blade and bus volume in the ramp-up stage that typically generate lower margins due to higher direct labor hours and material usage until those lines are at full capacity. General and administrative expenses for the quarter were $9.8 million, or 3.8% of net sales, as compared to $9.3 million in the same period in 2017, or 3.7% of net sales.
Before share-based compensation, G&A as a percentage of net sales was 3.2% and 3.3% in Q3 of 2018 and 2017, respectively. Net income for the quarter was $9.5 million as compared to $21.7 million in the same period of 2017. The decrease was primarily due to the operating results discussed above, partially offset by a benefit of approximately $11 million in Q3 of 2018 from the release of the valuation allowance against our U.S. net operating losses. This valuation allowance release is a result of our decision during the quarter to account for the impact of the global intangible low-taxed income, or GILTI, as a current period item. With this election and the adoption of ASC 606, we ceased to be in a cumulative loss position in the U.S., and we have positive evidence that our U.S. tax attributes will be fully realized in the future.
Diluted earnings per share was $0.26 for the quarter compared to earnings per share of $0.62 for the same period in 2017. Adjusted EBITDA decreased to $17.6 million compared to $27.9 million during the same period in 2017. Our adjusted EBITDA margin for the quarter was 6.9%, down from 11% in the third quarter of 2017. The decline was driven primarily by the startup and transition activity and the resultant loss volumes. Before startup and transition costs in both periods, our adjusted EBITDA margins were 15.3% and 15.9% in Q3 of 2018 and 2017, respectively. Moving on to Slide 12. We ended the quarter with $110.8 million of cash and cash equivalents, total debt of $133.7 million, and therefore net debt of $22.9 million, compared to net cash of $24.6 million at December 31, 2017.
The relatively small decrease in our cash position during the quarter demonstrates the strong cash generation capability of our business and our ability to fund the increased level of startup and transition activities, as well as significant growth-related CapEx. For the quarter, we had net cash provided by operating activities of $14.7 million, while spending $8.3 million on CapEx, resulting in free cash flow for the quarter of $6.3 million. Our balance sheet remains strong, and we continue to demonstrate the ability to fund our growth primarily with cash generated from our operations and the significant availability we have under current credit facilities. I'd like to update our key guidance metrics for 2018. Please turn to Slide 14. We expect total billings for 2018 of between $1.0 billion and $1.05 billion, while revenue under ASC 606 is expected to be within the same range.
We expect adjusted EBITDA for the full year to be towards the upper end of our $65 million-$70 million range. We expect to deliver between 2,420 and 2,440 wind blade sets in 2018. Blade ASP for the year will be in the range of $125,000-$130,000 per blade. Total dedicated lines at year-end will be between 51 and 55. Capital expenditures will be between $85 million and $90 million. Startup and transition costs will be between $74 million and $75 million, and net interest expense will be between $14 million and $14.5 million. I'd like to introduce our 2019 guidance. First, I would like to take you through some of the key drivers and assumptions. We expect to increase our dedicated manufacturing lines by year-end 2019 to between 62 and 65 lines.
Net sales growth of over 50% and adjusted EBITDA growth of over 85%, based on the midpoint of the guidance range. Significant planned investments in 2019 to drive growth in 2020 and beyond as follows. Lines in startup during the year of approximately 15, resulting in startup costs of between $30 million and $33 million. Lines in transition during the year of approximately 10, resulting in estimated transition costs or under-absorbed overhead of between $22 million and $25 million, and providing incremental potential revenue under contract of approximately $200 million. Eight of these lines are for GE in Iowa and Mexico. On average, we estimate the incremental EBITDA impact of a line in transition outside of the U.S. will range from $1.8 million-$2.7 million, depending on whether it is a minor transition or a full mold transition.
This is before considering transition fees that are paid by our customers or extensions to the terms of the agreements. This represents the expected lost contribution margin dollars on the lost volume during the transition, which generally takes one to one and a half quarters to complete. Overall utilization of the assumed 50 lines under contract as of the beginning of 2019 will be approximately 87% during 2019. Utilization is a new metric we're introducing today in order to help the investment community better understand the impact of startups and transitions on our net sales and total billings. Our calculations of utilization for 2019 are based on the assumption that we will start the year with 50 dedicated lines, and that any new lines signed during the remainder of 2018 and in 2019 will not provide net sales or billings until 2020.
Free cash flow in the range of $20 million-$25 million, along with the cash on our balance sheet, rigorous working capital management, and selective use of credit facilities when needed, will be used to fund our growth and development efforts and provide us with the flexibility needed to execute our strategy. We expect capital expenditures to range between $95 million and $100 million. Transitions and startups in 2018 and those in 2019 will drive an expected increase in ASP to between $135,000 and $140,000 per blade, up from the $125,000 to $130,000 in 2018. We expect that the continued conversion of our pipeline will necessitate additional facility or campus expansion during 2019, and we expect entering a new geography during 2019 to accommodate customer demand for another low-cost, world-class manufacturing location and to further diversify our global footprint and supply chain.
We plan to open a new tooling facility in Juárez, Mexico, to capitalize on the supply constraint for tooling globally and to take advantage of the infrastructure we already have in place in this low-cost and highly skilled labor market. We also expect to expand our tooling resources on a global scale to lift this constraint and assure execution of transitions and startups. We will continue to focus on day-to-day execution to continue driving down cycle times and direct labor hours and collaborating with our supplier base for raw material pricing, certainty of supply, and further innovation, with a strong emphasis on more efficient execution of startups and transitions to drive faster ramps and reduce overall volume loss and costs for each.
We will continue to use productivity and throughput improvements driven by the TPI Integrated Production System and our scale to drive our costs down and enable us to commit to customers' annual cost-out goals, provide them with more volume flexibility and enable them to remain competitive, gain share, and reduce LCOE, while at the same time protecting our margins through the shared gain feature in our contracts. Finally, we will leverage our investment in the automated pilot manufacturing line to advance our diversification strategy and expand the number of transportation-related production contracts over time. However, as in the past, the additional potential revenue from further diversification is not included in the 2019 guidance or 2020 targets. With that as a backdrop, please turn to slides 16 through 21 for our full guidance for 2019, as well as net sales, total billings, and adjusted EBITDA bridges.
For 2019, we expect net sales and total billings of between $1.5 billion and $1.6 billion. The range is above the target we provided last year. A bridge showing the major components of the net increase from 2017 through 2019 can be found on slide 18. Adjusted EBITDA of between $120 million and $130 million. Our range is below the midpoint of the target we originally provided, primarily as a result of the additional startup and transition costs over what was anticipated when the targets were developed. A bridge from 2017 through 2019 can be found on slide 19. Fully diluted earnings per share of between $1.24 to $1.35. Sets invoiced of between 3,300 and 3,500. Average sales price per blade of between $135,000 and $140,000. Estimated megawatts of sets delivered of approximately 9,800 to 10,400. Dedicated manufacturing lines at year-end to be between 62 and 65.
Manufacturing lines installed at year-end to be between 50 and 52. Manufacturing lines in startup during the year to be approximately 15. Manufacturing lines in transition during the year are expected to be approximately 10. Line utilization based on lines under contract as of December 31, 2018, of between 86% and 88%. Startup costs of between $30 million and $33 million. Transition costs of between $22 million and $25 million. CapEx to be between $95 million and $100 million, approximately 85% of which are growth-related. Our effective tax rate to be between 20% and 25%. Depreciation and amortization of between $40 million and $45 million. Interest expense of between $12 million and $13 million. Share-based compensation expense of between $9.5 million and $10 million. For 2020, we're providing the following preliminary targets. Net sales and total billings of between $1.7 billion and $1.9 billion.
This represents over 15% top-line growth from the midpoint of 2019 to the midpoint of this range and represents a utilization rate of approximately 80% for the 62 to 65 dedicated lines we expect to have under contract by the end of 2019. Adjusted EBITDA of approximately 10%, or between $170 million and $190 million. This assumes a more normalized level of transitions in 2020, given the heavy transitions we have had in 2018 and will have in 2019, as well as the push in the U.S. market to meet the installation and commissioning deadlines by the end of 2020 for the 100% PTC. It also assumes lines in startup in the low double digits. TPI is on track to more than double our net sales by 2021 from approximately $1 billion in 2018 and continue to move towards our target of 12% adjusted EBITDA over time.
With that, I turn it back over to Steve to wrap up, and then we'll take your questions. Steve?
Thanks, Bill. We're pleased with TPI's third quarter and year-to-date results, as well as the successful delivery of many of our aggressive startup and transition programs. We remain very confident in our global competitive position and the application of our dedicated supplier model to take advantage of the strength of the growing regions of the wind market, the trend toward blade outsourcing, and the opportunities for market share gains provided by the current competitive dynamic. We have clear line of sight to doubling our current wind revenue to more than $2 billion in 2021. In addition, we are pleased with the traction we are seeing in our transportation development programs. We are focused on finishing the year strong and are looking forward to an exciting and rewarding growth years in 2019 and 2020. Thank you again for your time today, and with that, operator, please open the line for questions.
Great. Thank you. At this time, we will be conducting a question and answer session. If you'd like to ask a question, please press star, one on your telephone keypad. A confirmation tone will indicate your line is in the question queue. You may press star, two if you would like to move your question from the queue. Please limit to one question and one follow-up question and re-queue for any additional questions. Also, for participants using speaker equipment, it may be necessary to pick up your handset before pressing the star keys. One moment please while we poll for questions. Our first question is from Philip Shen from Roth Capital Partners, please go ahead.
Hey, guys. Thanks for all the detail through 2020. It's really helpful. As we calculate the implied EBITDA margin for 2019 and 2020 now, I think we're at 8% and 10% respectively. How conservative do you think you guys are being on this? In the past, we've talked about kind of a target 12% EBITDA margin. Are we accounting for the startups and transitions here, and then in turn, what is the risk of higher than expected transitions that you have not factored in for 2019 and 2020? Are all the lines accounted for effectively by the time we get through this process so that maybe there's nothing left, or there are still some potential lines that could transition? I know there's a lot there, but I wanted to put it out on the table. Thanks.
Hey, thanks, Phil. You put a lot in one question. That's pretty good. Yeah. I think clearly, there's some conservatism in these numbers, but as we learned this year, sometimes the transitions can be unexpected. We think we've got a pretty good bead on 2019, on what our customers are asking for at this point from both a transition and a startup standpoint. For 2020, it's a little harder to tell, but for the reasons we talked about in the prepared remarks, given the heavy level of transition both this year and what we're expecting in 2019, the vast majority of our lines will either be at a very market-specific blade size, or that'll be a brand-new blade from a startup. We think for that reason, as well as the push for installation and commissioning in the U.S.
market in 2020, we think the level will be less than we've seen in the last couple of years.
Okay. Great. Then you guys had mentioned that you will likely add a geography sometime in 2019. Can you give us a little more color on this? Has it actually been decided and now it's just a matter of when both parties feel comfortable with the announcements? What characteristics might you be able to share about that geography and how many potential lines could be in that geography?
Yeah. Phil, it's Steve. We're not going to be specific about a particular location at this point. Consistent with what we've been saying for some time, though, as we think about further diversification geographically around the world, you can be certain that our next move would be another low-cost hub to serve multiple markets in a cost-effective way. We do like the idea of some additional expansion and some additional diversification through that. Most all the campuses that we would expect to start, we would start with probably a minimum of four lines, anticipate eight, and then, roughly, let's say eight, and then make sure that the geography, the land site, is capable of doing even more.
That's probably the best way for us to scope it for you today, and then we'll announce more as we're ready.
Great. That provides some nice color. Thanks, Steve. Thanks, Bill. I'll pass it on.
Thanks, Phil.
Our next question is from Joseph Osha from JMP Securities. Please go ahead.
Hi, this is actually Hilary on for Joe. I just want to ask a quick follow-up to the new geography, if you could just kind of give some thoughts on how you prioritize and/or balance the different growth avenues, either it be expanding with new customers, the existing customers, or to continue the geographic expansion.
Yeah. Hilary, we look at a balance of each of those. The pipeline that we'll convert on now that we've expanded through 2020 is a combination of current customers expanding more lines, as well as some new customer opportunities. It includes, as we've said, at least one new geography through that timeframe. It includes mostly onshore and potential of some offshore lines. It's just a balanced view of exactly that. We're very keen to continue to grow with the major players in the wind space. We want to make sure that we're matching up with strong customers in the regions that they're going to serve. We have a prioritized pipeline that on purpose, we set a cut line, and we rank order them the way we think it's best for us and them, and then decide what we're going to do, what makes the cut, so to speak.
That's the rationale we apply, and we've consistently applied that really for the last few years.
Okay, great. Just also thinking about growth, if you could kind of provide a little more color on how you think about either taking more share or looking to capitalize on that outsourcing trend that we've been seeing.
Yeah. I think the main driver for our growth has been, and continues to be outsourcing, but it ends up in our global share, actually, right? If folks are outsourcing from an in-house facility to a low-cost hub that TPI could have multiple customers spread overhead, drive cost, help our customers tap new markets, that is the main driver that's providing such a strong growth backdrop for our business. That is the main driver. The result of that ends up being the ultimate market share, and we're continuing to drive additional gigawatts each year through that mechanism.
Okay, great. Thank you.
Thank you.
Our next question is from Eric Stine from Craig-Hallum. Please go ahead.
Hi, Steve. Hi, Bill.
Hey, Eric.
Hey. Just interested in your commentary about the industry moving to families of blades. Just curious, as part of that transition, do you think about that as there's a blade length where people start to transition in that way? Or is this really just an OEM decision and when they decide that they're going to make that sort of a change? Because that's a pretty big change, I would think.
Yeah. It's actually an evolution, Eric, from ideas that have been implemented a little bit over the last few years to just sharpen the focus on it. The reason is to be able to add length to a blade through a more rapid tip extension rather than pulling out the full investment of the molds, which for us takes more time, more downtime, more EBITDA burn during the major transition, if you will. It's not really a brand new idea, but it is being implemented in a more orchestrated way. If you think about the good news about our market being so much driven now by economics and customer choice, that's really good news to where we're not talking as much about feed-in tariffs or government regulations over time, more about economics.
The trade of that is it's a battle for economics, our customers are all driving really hard for new product introductions to drive LCOE. If you think about if you were running their business, to run their business and return invested capital to your shareholders, you'd want to reduce the new product introduction costs. For them to just orchestrate a little smarter on families of blades, where it's easier to capitalize on various wind conditions and cheaper and faster to add length, that's just a smart move. It's a smart move for them, and it'll be helpful for us in terms of the speed and cost of transitions.
Got it. That's helpful. Then maybe just on the new investment, just want to make sure I understand it correctly. Is it right to think about that that's not necessarily an investment for a specific program, but it's more of a demonstration or a development facility that you would then incorporate in future programs?
Yeah, that's right. As we said, it's a pilot line. It's not an ultimate production facility. We expect to have some revenue capacity from it when we're done, but the purpose of it initially is to develop some technology for automotive type products, which means smaller physical parts, Eric, but it also means much shorter cycle times. These are parts that will be made in minutes, not hours. Wind blades weigh 15-20 tons, and it takes us 24+ hours to make them. These will be parts that'll be fabricated in a few minutes. We need to demonstrate the technology. We're using it to create both product intellectual property and process materials and process know-how IP, if you will. Then it's also making sure that we can demonstrate at-rate production to the point that our new customers would say we're ready to scale.
It's very important for us to demonstrate it fully, be ready to ramp, and then get production decisions made with our customers in a way that feeds our revenue growth at the right time. That's the objective at this stage.
Okay. Fair to say that is at least in part targeted at some current programs that you have.
Yeah, we have a handful of development programs as we've reported on before, primarily EV related. There's also, as we've mentioned before, a Navistar truck program.
a couple of other initiatives that may not be just EV driven. This is largely to penetrate the EV market opportunity for TPI. In the EV space, that's where high strength and lightweight, we believe offers even more value, where weight savings equates to added range. There are multiple customer programs that could be supported from this pilot line. We're working on them already. We've described those in the past. Yes, think of it as generic process equipment and technology that could support multiple programs.
Okay, thanks a lot.
You bet.
Our next question is from Paul Coster from JPMorgan. Please go ahead.
Yeah, thanks for taking my questions. Couple of quick ones. It seems like you've got 23 lines in your pipeline. Are these likely to include new logos or is it the existing customer base? In two years from now, should we expect your revenues to be more diversified than they are today? Just within the winds context.
Yeah, Paul. Our pipeline, the 23 molds, is made up of existing customers and a couple of new customers. Yes, we would expect to add one or two logos over time. We also expect to grow with some of our current customers. It's a combination of both of those as we go forward, new geographies, and then as I think we said, a combination of mostly onshore, but a little bit of offshore. All of those are against this goal of kind of really solid growth, but diversified as well. Yes, we would expect further geographic diversity, a little more customer diversity, and a bit of both onshore and offshore diversity as well.
I imagine it's more than wishful thinking that you've got a couple of new logos lined up. Can you characterize the nature of the relationship with those potential customers as things currently stand? Then my one other question is, as we roll from 2018 to 2019, given the fact that you'll be adding a new geography, the CapEx just didn't look like it was bumping up in the manner I would have expected. Perhaps you can just comment on that as well. Thanks, Steve.
Yeah. I'll take that. I don't think we'll speak specifically about the nature of the relationships with the potential new logos. On the CapEx side, Paul, it actually bumps up a bit. If you think about a new geography, the first year of it, if we think about 2019, a fair portion of that will be a real estate project that doesn't sit on our books but sits on the developer partner's books. We'll have some CapEx towards the end of that as we begin to put CapEx into that building itself. This CapEx for 2019 is some carryover from 2018 related to our new plant in Yangzhou, as well as in Matamoros, Mexico, as well as some additional CapEx that we'll be doing in some expansion in our existing Mexico facilities as well, Yeah. That's it.
Yeah, Paul, I think you can imagine in a brand new location, the first year or so is mostly a real estate project, and then a lot of our CapEx spending follows as well. Not all of the CapEx for that new location would be in calendar 2019, actually, or not necessarily so. To Bill's point, in terms of the relationships with new customers, what I guess I'd say generally is there are a limited number of targets for us in terms of customers that are really important targets for TPI. You can imagine we all know one another, and the relationships are developed to the point that we've all been discussing ideas about what to do or not to do for some time. The key really is not so much to develop the relationship from scratch, but it's to close. Right?
If we close on deals where a customer says, "Let's go," and we say, "We're ready," then we convert a customer on the pipeline to a new logo in our decks. It's more about converting, I would say, than introductions, if you will.
Got it. Thank you.
You bet. Thanks, Paul.
Our next question is from Pavel Molchanov from Raymond James. Please go ahead.
Thanks for taking the question. As I think about your non-blade sales, they've been running at around $20 million per quarter or so for the last three quarters. Given that your overall revenue guide for next year is up 50%, should we assume that non-blade revenue increases at a faster than average rate, or should it be more in line with the overall 50%?
Yeah. Our guidance for next year is $115-$120 of non-blade, if you will. Right? It's a pretty good increase over this year. We looked at about $85 million this year. $80 million-$85 million this year, Pavel?
Right. I'm curious why it wouldn't grow faster than wind, particularly given that the Newton, Iowa facility only opened in the middle of 2018, right?
Yeah. Pavel, it's Steve. I think one thing we might just try to remind you of in terms of how we're developing this business, the mission is to build a half a billion-dollar transportation-related revenue over the next handful of years. We're thinking about making sure that we're a growth company in years four and beyond, if you will. The development of the transportation segment, as we've shared, is going to take some time, and that's okay. We've got plenty of wind-related short-term growth. I just don't think It's going to be one win at a time. It's going to take some time for that to develop. It's probably, in our case, not so much just scaling as to whether the growth rate is equal to greater than, less than wind.
It's really more one of, if you think about the pilot line we just spoke about, we're going to make an investment. It's going to take time for that to come up. We'll develop technology. We'll work to convert customers. All that in this transportation automotive space, it takes some time. I think you could probably expect that it will take us some time.
Okay. Can I just get a quick comment on epoxy resin pricing and what's the latest on that source of input inflation?
Yeah. We've just recently been through our latest round of pricing for 2019, and on a per kilogram basis, our price is down year-over-year as it was in 2018 compared to 2017 and 2017 compared to 2016. Based on our relationships and our supplier arrangements, we're not seeing any impact on resin for 2019 moving forward, notwithstanding spot prices. The spot prices have stabilized a bit over the year.
Okay. Very helpful. Thank you, guys.
Yep.
Thanks, Pavel.
All right. Once again, as a reminder, if you'd like to ask a question, it is star one. Our next question here is from Chip Moore from Canaccord. Please go ahead.
Good evening. Hey, guys.
Hey, Chip. How are you?
Hey, Chip.
I'm good. On the startup and transition costs for 2018, just wondering if you can talk about a slight tick up versus the prior view with two fewer lines in transition. I think you talked about a customer being delayed a little bit, but just if you could provide a little more color on that.
Yeah, sure. It did tick up a little bit. The way we think about startups and/or transitions is we leave them in the startup phase or the transition phase until we get up to a line rate of at least 90% of capacity. As a result of a few of the delays we've had during a couple of the startups and transitions, that's the reason for the tick up. The startup costs will tick up if we have a little bit of a push in the time. That's the primary reason for it.
Got it. Was any of it related to the tooling when you talked about the new facility in Juarez?
No, not really. I think we've already started building tooling out of Juarez. We're going to move it into a new facility here in the first half of 2019. That was done. We talked about the highly skilled labor force in Juarez, and we have a great team down there. That was done very seamlessly. There was no traditional startup cost, if you will, in there. We just picked it up and went pretty quickly.
Okay. Just one last one from me. You talked about tariffs and the potential for modest impact. Are you contemplating any shifting of blades at all in the outlook?
Yeah. I think, Chip, the comment that we made there is that is really a question for our customers in that we build the blades and put them outside. They pick them up, and they deal with logistics and decide where in the world the various blades are going to be shipped to. The point we have made is, even though there could be a cost uptick for blades coming to China, if the tariff were to stay in place, the demand on our China operations is still strong, but our customer may just choose to move the blades to another location. 10% is a big number, 25% is even a bigger number. I doubt at 25% that a lot of blades come from China to the U.S. or other components unless they have to.
Our point there is the global footprint is a strength and provides flexibility in that perhaps more blades get pulled from Mexico, for example, for the U.S. market, and more of the blades made in China might go elsewhere. For 2019, the demand profile is very strong on China, even though there is likely to be a 25% tariff for some period of time. We are just moving past it a bit. Supply chains adapt. When you are global as we are, we are able to adapt along with that. It would be a shame from our perspective. We support free trade. We think the tariffs are a bad idea.
It would be a shame if the cost reduction that we have been able to materialize in the U.S., 67% over the last eight years, if that were to go up by 5% or 10%, that would be going in the wrong direction from what would be desired. I think reality is you will see some shifting of the supply chain and mitigation of the cost increases over time.
Perfect. Appreciate it. Thanks, guys.
Thanks, Chip.
Thank you, Chip.
Our next question is from Jeff Osborne from Cowen and Company. Please go ahead.
Hey, good afternoon, guys. Maybe just a follow-up on the China question. Can you talk about where the blades are going outside? It sounded like 20% of the blades produced in China are going to the U.S., but where's the other 80% going?
Yeah, Jeff, in a given period, the blades will go to various countries that our customers choose. We've had examples where blades go to India, they go to Australia. They'll even go to pockets within Europe or South Africa, various markets. Once the blades are on the water, they can be moved quite some distance without much additional transportation cost. I think the point for us is we don't make those decisions, our customers do. The question is: Is the demand on our factory still strong? Meaning, is their volume around the world strong enough to where they would pull from China and perhaps ship the blades to another location? That is what's happening. It's different countries, depending on where our customers win projects, what their other footprint options might be. They make the moves based on what's most cost effective for them.
What we care about is to make sure the demand pull on each of our operations remains really strong, and that's the case. Even with the tariffs, that's still the case.
Makes sense. I appreciate the detail there.
You bet.
A few more from my end, if you don't mind. On the transition timing, can you just give us a sense of perspective across the four geographies you have? Is that a two-quarter process, three-quarter process? It sounded like it was going a little bit slower than anticipated. I assume just with the massive growth you've had, that maybe that's a challenge or maybe customer molds versus your own molds. I'm not sure what the exact issue is, how should we think about handicapping the EBITDA as it relates to specific transitions that you've called out?
We really haven't called out any specific ones, in general, Jeff, they'll take any-
I think you called out GE. Sorry to interrupt, I think you did call out GE actually-
Yeah
in 2019.
Yeah. As an addition, right?
How long do you think it'll take Iowa and Mexico to transition? Is that a six-month process or no?
Generally, it takes a quarter to a quarter and a half to fully transition. These are minor transitions, tip transitions, if you will. It generally takes a quarter to a quarter and a half from the time we stop production until we get back to full capacity on each mold. That's the general time. Yep.
Yep. That's helpful. Then I guess I was surprised you alluded to the Iowa transition on the last call, which is certainly an endorsement of the competitiveness of wind in the Midwest, in an ex PTC world. I guess I was expecting now that you've included the GE transition in Iowa in your guidance, I was anticipating that, given you'll be down for a quarter to a quarter and a half, it sounds like, and you only have four or five quarters left. I guess six or seven, sorry, on the contract. How do you just think about the return on investment of that downtime versus changing the blades? I was anticipating some type of press release today of GE extended the contract for a year or two.
Is that something that still could be had, or should we think about the Newton facility with the six lines going to five closing still at the end of 2020?
Yeah. Jeff, it's Steve. I don't think we're at a place where we could comment on one specific customer kind of as specifically as you're asking. I think in general, it's true that if our customer invests in transitioning to a new blade model, they're making an investment in that, as are we in terms of an EBITDA dilution during that transition period. Our customer and ourselves, you can imagine we would both have to see kind of reasonable runway to make the return on invested capital make sense. Your question's right. That point is correct. Individual decisions will be based on the individual customer and individual factory location for TPI. I don't think we can be more specific about that one location, except to say that is the right way for us to the way we think about it.
Got it. The last one I had is just, it was unclear to me on either on a rule of thumb basis or for 2020 in particular, how many lines are you calling out for transition in that period? There'll likely be a surge in demand, which I agree with. You're a global company. I just wasn't sure with the other facilities if there's a sense of we should be modeling if you have 60 to 65 lines in operation, that 10%-15% every year in transition or any back of the envelope math or if there's any specific commentary about the 2020 methodology you have, that would be helpful.
Yeah. We talked about the fact that of those 60 to 62 to 65 lines, we thought we'd be at about 80% utilization of those lines during 2020. That can give you a little feel that there will be some in we think there could be some in transition. Again, in 2020, we think that number is pretty low for the reasons we stated. We also talked about low double digits of lines in startup during the year. That's about as specific-
I get it
as we can get at this point.
Jeff, as Bill said, that 80% utilization number is maybe a little simpler way for you guys to be thinking about how to model some of this rather than getting every transition or every startup perfect in terms of timing and overall impact. Those things move around a little bit, right, as we've shown. The 80% utilization is probably the right way to think about it.
Yeah. Maybe the last one, and I don't know if you can be as specific as what I'm hoping for, but you've got 51 lines in operation. Is there any perspective you can offer, stripping out GE, which is about to be in transition, how many of the blade designs that you're producing are over two years old? Is there a way to look at the lifetime or the life cycle of a blade?
You've had some new customers, you've had some existing customers, you've had a lot of changes, it would be helpful if just given the dynamic nature of the market moving to auctions, if there's a way to look at the average life of a blade, because that seems to be getting shorter and shorter over time, and that's certainly an investor fear that we hear from folks about just that these tip changes and blade changes will be in perpetuity and not just one-time items.
Yeah.
I don't know if there's any data that you can provide that would ease that concern.
Jeff, I don't think we can get more specific on the exact number of lines that are over two years, for example, today. What we have tried to say, and I think it's a direction that will prove to be true, it's going to make sense for a bunch of reasons, is that the transitions will slow over time, and they kind of need to, for a combination of reasons. The other truth is we are getting faster at the transitions, and our customers, as we said a few minutes ago, are doing more in terms of families of blades to make more of the transitions be tip and minor related than major related. All that's kind of moving in the right direction to help mitigate the impact.
The other thing is when blades get to a particular length and it's no longer going to be cost-effective to truck the next bigger model, that's a reason to not go to that next bigger model. If it's a split blade, some of the new modular blade ideas, that'll make sense perhaps to move at that point to a modular blade. Which, by the way, we could build in the same factories as we build the larger blades. In fact, they'll be smaller pieces and easier in a sense, and quicker to transition the small part mold. All this is in a direction to make it faster, cheaper to transition, and then we'll see some inertia on the transition. That's a general answer. Not as specific as you might like, but that's the direction that we see.
Great. I appreciate it. That's all I had.
Thanks, Jeff.
Thank you. This concludes the question and answer session. I'd like to turn the floor back to management for any closing comments.
Thanks, operator. Thanks all of you again for your interest in TPI Composites. We look forward to continuing to update you on our progress. Thanks again.
This concludes today's teleconference. You may disconnect your lines at this time. Thank you for your participation.