Exhibit 99.1
MathStar, Inc.
Conference Call Script
Quarter ended June 30, 2006
Conference call date: August 10, 2006
Conference call time: 8:00 AM PDT
Call Duration 60minutes
Welcome to the MathStar, Inc. quarterly earnings conference call for the quarter ended June 30, 2006. All lines have been placed on mute to avoid any background noise. After the presentation there will be a question and answer period. Please limit your questions to one and refrain from multipart questions in order to give management time to answer as many questions as time permits. If time permits, additional questions will be taken. I would like to reference you to the forward-looking statements and risk factors disclosed in MathStar’s Annual Report on Form 10-K for 2005. With that, I will turn the call over to Doug Pihl, MathStar’s Chief Executive Officer.
Good morning. Thank you for joining us for our 2nd quarter 2006 conference call. With me today are, Jim Cruckshank MathStar’s Chief Financial Officer, and Dan Sweeney, our Chief Operating Officer.
MathStar’s second quarter 2006 was pivotal for our company’s goal of becoming the segment defining Company in high-performance reprogrammable logic. I am pleased to report that we achieved major milestones that are key to the successful launch of our new FPOA technology into commercial markets.
First, as we announced last month, we achieved our most significant milestone: we taped-out the production version of our FPOA chip. As you know, it was a difficult and complex process to bring the commercial version of the chip up to the exacting standards needed for high volume production. The production version of the FPOA incorporates improvements to the development version that were suggested by our early adopter customers, making the FPOA ideal for our targeted application markets. We also made numerous adjustments to assure optimal “timing closure,” and, after exhaustive testing and re-working, we are confident that our design will achieve production yields in our target range of over 70%. This is important to us because it will help to assure the high margins that we have been promising. Our production FPOA contains 400 silicon objects and operates at 1 Gigahertz, making it possible to execute 400 billion operations per second. The FPOA continues to lead the high-performance market: it is 2 – 4 x faster than the nearest competitive programmable logic products.
At tape-out, the FPOA was sent to Taiwan Semiconductor for fabrication, and it is in the manufacturing process now. We expect to have silicon back for internal verification by the end of September and will begin shipping commercial FPOAs to lead customers in October. Full production shipments will begin in December. FPOAs will be available in a range of speeds from 400 Megahertz to 1 GHz, with the fastest chips priced in the range of $250-$300 per device in thousand unit quantities.
While the tape out of the production version of the FPOA was, of course, our primary focus in the quarter, we made excellent progress in a number of key areas:
· Our customer pipeline is strong and growing. We completed the build out of our manufacturers rep organization in North America and added our first representative in the UK. These sales representatives are focused on the high-performance segment of the market, and they are eager to offer a 1 GigaHertz field programmable chip to their customers for the first time ever. They are also looking forward to having commercial versions of the FPOA to deliver to prospective customers for their evaluation. As of today, MathStar has trained over 100 sales reps on the FPOA. Our three Area Sales Managers accompany reps on sales calls and help to prioritize a growing list of opportunities. Later, Dan will explain in more detail how we are progressing with customer engagements.
· We are attracting prospective customers in a range of markets, but continue to focus on the machine vision and high performance video markets. We believe we can rapidly penetrate these markets because they compete on the basis of high-performance, and we can help our customers to provide the fastest products in their respective markets. A majority of our first commercial shipments will be to customers in the machine vision and high performance video markets, and these customers have committed to creating product families based on the FPOA. We will continue to penetrate these initial markets as well as markets such as military, security surveillance, medical imaging and test and measurement. We believe the FPOA is ideally suited for a wide range of applications and we will build on our strengths in early markets to drive growth in adjacent markets and beyond. We have been especially pleased with the response to our offerings in the high performance video space as we believe video-related markets will grow and expand as companies and consumers process and communicate more and more through video images.
· We are releasing exciting, industry-leading application libraries for our key target markets. In the 2nd quarter, we released the first application libraries in Machine Vision and Professional Video. These application libraries greatly facilitate our sales process by showing prospective customers real functions running on our chip. Seeing a real application that prospective customers understand enables them to do a faster evaluation of our chip and gain a better understanding of its capabilities in their product space. These library elements will also enable a faster time to market for our customer’s products by using standard algorithm code they license from us. The programmable nature of our chip also allows customers to customize our algorithms and add their proprietary code.
· One example of our industry-leading applications is our JPEG2000 Encoder. JPEG2000 is an industry standard video compression algorithm used by many of MathStar’s customers in target applications. We introduced this library module in July and it will perform at the rate of 200 million pixels per second. This will enable encoding in high performance video and machine vision applications at rates of over 200 frames per second.
· In addition, we engaged with two independent, third party IP providers to develop additional library elements and expand our libraries faster than if we did all the development in-house. As we expand our family of third party providers, our customers will be able to access these qualified FPOA developers to assist in FPOA programming.
· In May, we released Version 2.0 of our FPOA Design tools that our customers use to design, simulate and debug applications to be used on the commercial version of our FPOA. This release is the next step in our tools development to make FPOAs easier to program and speed time to market for customers using FPOAs in their products.
· Our Marketing team is greatly expanding industry and customer awareness of the FPOA. If you have not already done so, we encourage you to read the InStat report on the FPOA. InStat is a highly respected third party research firm, and their report appeared in their well-read industry publication. You can read it at our website. We believe it is a valuable piece for educating prospective customers. We are planning an extensive media campaign as the commercial FPOA is launched this fall, so please continue to visit our website for further third party news and reports.
· We completed the product definition of the next two FPOA devices to be introduced next year. These extensions to our family of FPOA chips will offer customers a wider range of features and benefits to meet a wider range of customer demand. The first will be taped out in the first half of 2007 and the second by the end of 2007. These products will be implemented in the 90 nm process technology and will offer even higher performance and lower power.
We believe we are well positioned for rapid revenue growth as the production FPOA is launched into markets demanding a high-speed, reprogrammable device. Jim will now provide a financial update followed by Dan to provide more details in his operational update. Jim…
Thanks, Doug.
For the three months ended June 30, 2006, our revenues were $10,000 compared to $25,000 for the same period last year. Revenues for the three months ended June 30, 2006 were from the sale of software licenses and non-recurring engineering fees, and revenues for the three months ended June 30, 2005 were from the sale of FPOA development kits.
Our research and development expenses increased $1.3 million or 70% to $3.2 million for the three months ended June 30, 2006, from $1.9 million for the three months ended June 30, 2005. The increase was the result of increased payroll costs ($0.5 million, which includes $0.1 million of compensation costs related to stock options and restricted stock), increased outside engineering and design consultants ($0.3 million), increased engineering materials and electronic design software tools ($0.2 million) and increased overhead costs ($0.2 million, due to an allocation of certain overhead costs). The increase in payroll costs and the use of outside engineering and design consultants is the result of increased staffing and efforts to complete the FPOA chip design and to continue development of our library of development algorithms for use by our customers.
For the three months ended June 30, 2006, our selling, general and administrative expenses increased $.8 million or 51% to $2.3 million, compared to $1.6 million for the three months ended June 30, 2005. The increase was primarily the result of increased salaries ($0.6 million, which includes a net increase of $0.5 million for charges related to non-cash compensation costs of stock options and restricted stock), as we hired a chief financial officer and a controller and other staff in the second half of 2005. Other expenses increased $0.3 million. These increases were offset by an allocation in the three months ended June 30, 2006 of certain overhead costs ($0.2 million) to research and development.
Turning to the six months ended June 30, 2006. We generated revenues of $18,000 compared to $40,000 for the six months ended June 30, 2005. Revenues for the six months ended June 30, 2006 were from software license fees and non-recurring engineering fees, and revenues for the six months ended June 30, 2005 were from the sale of FPOA development kits.
For the six months ended June 30, 2006, we incurred research and development costs of $6.0 million, an increase of $1.7 million or 39% over research and development costs for the six months ended June 30, 2005 of $4.3 million. The increase was the result of increased payroll costs ($1.1 million, which includes $0.1 million of compensation costs related to stock options and restricted stock), increased outside engineering and design consultants ($0.7 million) and increased overhead costs ($0.5 million, due to an allocation of certain overhead costs). These were offset by lower engineering material costs ($0.8 million, as we incurred the cost of taping out a chip in the six months ended June 30, 2005 and did not tape out a chip during the six months ended June 30, 2006). The increase in payroll costs and the use of outside engineering and design consultants is the result of increased staffing and efforts to complete the FPOA chip design and to continue development of our library of development algorithms for use by our customers.
Selling, general and administrative costs increased $1.8 million or 68% to $4.6 million for the six months ended June 30, 2006, from $2.7 million for the six months ended June 30, 2005. The increase was primarily the result of increased salaries ($1.7 million, which includes $1.4 million of compensation costs related to expensing of stock options and restricted stock), as we hired a chief operating officer, chief financial officer, vice presidents of marketing and engineering and a controller and other staff in the first half of 2005. Other expense increased $0.7 million. These increases were offset by an allocation in the six months ended June 30, 2006 of certain overhead costs ($0.5 million) to research and development.
Our cash used in operations and investing activities for the six months ended June 30, 2006 was $8.6 million compared to $6.7 million for the same period last year. The $1.9 million increase was primarily the result of increased R&D expenditures, building the SG&A infrastructure and costs associated with being a public company.
Our cash balance at June 30, 2006 was $11.8 million. The only significant increase in burn rate during the balance of 2006 will be the one-time cost associated with taping and tooling for our first production FPOA of approximately $1 million that will impact us in third quarter. Based on our cash balance, cash burn and our expectations about the timing and extent of chip revenue during 2006, we believe we will have enough cash to carry us through 2006. Based on our discussions with investment bankers, we believe we will complete an additional financing before year-end. Although we have not finalized on the structure, we believe it will be a private equity placement. We will disclose additional information when appropriate.
We anticipate being Sarbanes-Oxley compliant well ahead of the regulatory requirements by the end of December 2006. I am also very pleased to report that we anticipate being ISO 9001 certified later this year. This is critical for closing design wins at tier 1 customers. .
I will now turn it over to Dan for an Operational update.
DAN
Thanks Jim.
Before I get into the operational details of the presentation, I do want to emphasize the level of effort that went into ensuring a quality tapeout of MathStar’s first production level FPOA. We made the difficult decision to delay tapeout 30 days to ensure that the defect density of TSMC’s 130nm process and not design marginalities in the FPOA would limit production yields. As Doug mentioned we successfully taped out on July 28th and expect 1st silicon by the end of September and provide customers initial samples in October of this year. We expect to ship full production silicon in December.
MathStar has also worked to ensure unconstrained availability of the FPOA as it ramps into production. Two initial lots of wafers were ordered at tapeout. Typically, the first chips from these wafers will be packaged and tested as complete assemblies. Although there is a small cost associated with packaging these untested parts, the overall schedule is faster this way. Packaging substrate design is complete and substrates are on order to support the steepest anticipated production ramp. The production test plan is complete and we expect to hit our target yields and test times, which will support target product margins. We are confident that production FPOAs will be available to support our sales ramp.
During the first half of the year, MathStar focused on bringing up a manufacturer’s rep sales infrastructure across North America. We’ve begun bringing up Europe with the first reps contracted in the UK. These reps provided MathStar instant access to system manufacturers in our target market segments. We’ve generated over 500 leads in the first half of the year to feed as opportunities into this rep network. We believe we will have generated over 1000 leads by year-end. We are now tracking over 40 design win opportunities in various stages in the design win funnel. We are adding 2-3 qualified design opportunities every week and are confident that we will have nearly 100 qualified design opportunities in the pipeline by year-end. Management remains committed to a minimum of 25 design wins but delays that pushed customer sample availability into early Q4 make this goal more challenging. Over 90% of the opportunities are in the target markets of Professional Video, Machine Vision, Medical Imaging, Military/Aerospace and Test & Measurement.
Software is key to MathStar’s long-term success. We are investing in two critical areas, the software tools our customers use to program the FPOA and FPOA application libraries used as building blocks by MathStar customers to lower their development costs and speed time-to-market.
In the Tools area, we made significant progress in the second quarter of 2006 towards maturing the tools our customers use to design and program our FPOAs. Tools development efforts in the second quarter of 2006 supported the announced release of Version 2.0 of the FPOA Design Tools in early May. Version 2.0 of the FPOA Design Tools includes the FPOA Connection and Assignment Tool (COAST Version 2.0), and Visual Elite™ version 2005.2 from Summit Design™. These tools allow customers to design, simulate and debug applications targeted for the next generation FPOA today prior to actual silicon availability in October.
MathStar is pro-actively developing a core of application libraries to support customers in targeted FPOA markets. In late April, MathStar announced the Professional Video Library for the Field Programmable Object Array (FPOA). The first intellectual property in the Library is an MPEG2 Multi-Stream Decoder that can decode multiple standard definition streams or a single high definition stream. We introduced our initial cores for the Machine Vision Library in May, the color space conversion and flat field error correction cores both used in almost every machine vision implementation. As Doug mentioned earlier, in July we introduced the world’s Fastest Reprogrammable JPEG 2000 Encoder for the Field Programmable Object Array. JPEG2000 is used across MathStar’s target markets as a way of compressing video in both lossy and lossless formats.
During the second quarter we finalized agreements with two 3rd party developers for development of additional library cores to run on MathStar FPOAs. Not only will these 3rd party developers accelerate the library of available library cores, they will also be trained on the FPOA architecture and tools for customers who prefer to outsource their FPOA software development.
As we have been completing the tapeout of the current chip in Minneapolis, the Oregon silicon design team is fully staffed and well into the design of our next generation of FPOAs. Our current plan is to transition the silicon manufacturing process from 130nm to 90nm, which will lower our cost and power per object by up to 40%. We also plan to leverage the ability to vary the object mix and location in this next generation of FPOAs, producing 2 chips optimized for the needs of our target markets. We expect to start sampling the first of these two chips middle of next year and the second chip will sample before the end of the year. These new FPOAs utilize our current object designs converted to 90nm and purchased I/O already demonstrated on TSMC’s 90nm process, which gives us high confidence in meeting committed delivery dates.
We are well on the way to completing the transition from an R&D organization to a ISO-9001 compliant commercial enterprise. We are confident that this new organization will be able to meet the company goals for the 2nd half of 2006 and beyond.
I will now turn the call back over to Doug.
Thank you, Dan
MathStar has a clear opportunity to redefine the high performance segment of the Programmable Logic market with our 1GHz, reprogrammable FPOAs. We continue to believe that the FPOA offers the best solution for customers seeking high-performance in a reprogrammable, off-the-shelf device, and that the FPOA represents the next generation of programmable logic. We are focused on rapid, profitable growth that will create shareholder value and believe we are in the best position in our history to deliver on that promise.
That concludes our formal presentation. We have about 15 minutes remaining for questions.
