‘There are multiple stages of death’ – When an exit isn’t all it’s cracked up to be

Pearse Coyle was CEO of NVMdurance between 2012 and 2017, bringing a team of inventors to a stage where their innovative flash memory technology was acquired. But Coyle thinks it could have gone so much better.

Between about 2001 and 2011, NVMdurance developed a technology that was expected to revolutionise flash memory storage. The claim was of a 10x lifespan improvement to memory, a dramatic promise and one which caught the eye of every major flash memory manufacturer in the world.

Paid trials, millions of euros of investment and countless false dawns emerged before an acquisition in 2017. However, the difference between the highs and the lows were stark. At one stage, a €300 million + exit was on the horizon, at another stage, any exit was beyond the founders’ wildest dreams.

'It really felt like we were at the races, a couple of times'

Origin Story

Inventors Joe Sullivan and Conor Ryan had worked out a way to make a piece of flash memory that would normally age and lose functionality after 5,000 read-write cycles, now retain its charge for 50,000. In a similar way to how batteries lose their charge eventually, this would mean huge things for any industry leader that adopted the technology.

The duo had been working on this problem for a long time, “probably too long,” recalled Pearse Coyle, former CEO of the now-acquired business.

When it came to turning the idea into a business, and the business into a potential purchase, they found out too late, that they were too late.

“It really felt like we were at the races, a couple of times,” said Coyle. Looking back, it’s easy to see why.

Early on, Coyle had been working with a US investor called New Venture Partners. It had previously invested in this space, so he tapped them up and asked how they went about commercially assessing that opportunity.

They pointed him towards a man in San Francisco called Tom Burniece, and pretty soon the startup was motoring.

“Tom was very well connected. He went to all the key companies, I mean the ones at the very top of the food chain, and said ‘I have a bunch of guys who claim they can do ‘X’, is that of interest?’

“The response, almost universally, was: ‘We don’t believe they can do it, but that is really compelling. If that’s true, we would like to know about it’.”

Soon after, a major semiconductor manufacturer signed up to a trial worth tens of thousands of euros to NVMdurance. “We thought, ‘Oh, that’s quick’. We knew we were on to something.”

More paid trials emerged, again with tens of thousands of euros attached, more interest swirled, investment loomed.

The same US investor was interested and – thanks in part to NVMdurance putting significant effort into its attendance at the industry’s major annual trade show in the US, the Flash Memory Summit in Santa Clara, where it won the most innovative technology of the year – many more companies picked up the phone.

Following NDRC investing in NVMdurance, the company raised €250,000 in seed money – it would later add more than €2 million in Series A. The quick, major exit eluded them, however.

Stage of Death: Slow start, nothing to show for it

In 2014, three similar companies were acquired for a little more than €1 billion combined. Apple, SanDisk and Western Digital each acquired businesses much like NVMdurance. These deals were positive, in that it validated what Coyle and his team were doing, and negative, in that it ended NVMdurance’s paid trials that were ongoing at the time.

Interest remained in NVMdurance, though, with access to the top companies in computing and memory meaning that, on several occasions, Coyle, Sullivan and Ryan thought their massive exit was only around the corner.

Sadly for the trio, and the staff of up to 15 at one point, all based in Limerick, it wasn’t to be.

“Flash memory is the fundamental component in the tech business. Almost everybody making any gadget has flash memory in it. They could all look at extending the lifetime but we felt we were relevant at the very top of the supply chain. We got access to those top companies and that got investors excited,” said Coyle.

However, one major difference between NVMdurance and the trio of newly acquired startups, was that Coyle and his team had no tangible product. They had a process, but until it was commoditised, it was still unproven.

“The guys spent too long in incubation before spinning out of NDRC. Ultimately what killed us was we were a bit too late to the market,” said Coyle.

“First, there was some IP leakage from the inventors’ original work. Prior to NVMdurance, one of the companies they consulted for had taken portions of their idea and built it into their controller business. That company was later acquired for hundreds of millions of dollars.

“This led to one of our trials being cancelled, but we thought it validated what we were doing.”

One of the other major acquisitions hit them in a similar way. Trial ended, validation underlined.

'We hadn’t aligned the test verification processes, that nearly killed us'

Stage of Death: When trials go bad

“We were essentially going to the big guys with a piece of IP. We weren’t making our own version of flash memory that lasts longer. That would need never-ending capital, huge manufacturing and distribution structures. A Limerick-based startup just could not finance that.

“What the other guys had done was take a route to market by making flash memory controllers, which was a far less capital-intensive route to market. It was a way to commercialise their IP and sell completed products.”

By the time those rival acquisitions ended NVMdurance’s trials, they met with other companies and secured two more trials – one in South Korea, one in the US.

The first trial was a roaring success, but only in Limerick. While the team felt they proved the technology was effective, their client could not replicate the findings.

It had taken 18 months, and team had nothing to show for it. “We hadn’t aligned the test verification processes. That nearly killed us.”

The second trial, luckily, began later so the NVMdurance team could apply their learnings and make sure the same mistake didn’t happen again – and they didn’t. NVMdurance finally had the proof needed. The trial company was satisfied and paid a significant six-figure sum for a warrant (the right) to acquire a portion of the company at the next funding round. However, as was the theme with the company’s journey, the enemy was time.

“NAND is a commodity product and it is cyclical and, as our funding round approached, the trial company, one of the major memory makers, themselves were undergoing upheaval and a severe round of redundancies at home in the US. It was unpopular with the tech teams but there was no money for investment or acquisition,” said Coyle.

'It always looked like we were close to a big deal'

Stage of Death: Customers caught up

Throughout NVMdurance’s evolution, the lingering question of whether or not the technology worked remained. The trials were imperative and, generally speaking, positive. Until they weren’t.

The company wasn’t built to relentlessly raise money. The investors viewed it as a quick acquisition play. The founders viewed it as a quick acquisition. Coyle viewed it as a quick acquisition.

The market viewed it otherwise.

“I’m applying hindsight to the logic,” explained Coyle, “but at the time it always looked like we were close to a big deal. We were getting interest from an appropriate level.”

The problem was the companies most likely to acquire, did. And NVMdurance wasn’t one of the companies acquired. Also, as the company evolved, its IP – which could originally improve flash memory 10x – was now down to maybe 3x after the industry had caught up somewhat. That is less attractive to potential buyers.

Stage of Death: Failed investment

As a result of the final trial failing to result in a commercial outcome, investment opportunities dried up. One of the many badges of respectability for NVMdurance had been how it wasn’t just local, Irish investment propping it up, US investors had become involved. But, even armed with that, they struggled to get deals done.

“We had an idea for a pivot, and some traction for it, but we needed investment. Our existing investors were willing to follow through, if we could get a new investor. Most Silicon Valley investors said ‘no’. Then an Irish VC investor said they would, all but officially. However, three months later, they changed their mind. They had changed their internal deal-approval processes. We later found out that they never should have said they’d invest in the first place.”

This was the final blow.

'As we came to the deadline for offers, one by one they fell out of bed'

Exit Music

How did the exit come about? “Painfully. Painfully is the only word. There are multiple stages of death.”

Coyle sought to sell the company, and they got what he terms a “half-hearted acquisition bid” from a company in Finland. This was a blessing, as it meant Coyle could repackage what would look like a fire sale into what was actually an acquisition approach.

“To our surprise almost all the major memory and storage companies engaged with us. They all said they would like a serious look. We were surprised, we backed off trying to do it by ourselves and appointed a banker to run a serious sales process for us.

“Several companies did serious due diligence. We thought, ‘Wow, we will sell, it’s just a matter of price, don’t fold up the tent yet!’

“But, as we came to the deadline for offers, one by one they fell out of bed. No offers came. Then the offer of an acquihire was rebuffed by them all. More than one of them said that, as we were based in Limerick, Ireland and they had no flash memory dealings in Limerick, they couldn’t integrate a team.”

The team tried a couple of other ways but ultimately the signs on the wall became far too big to ignore and they made arrangements with shareholders to go through a winding down process of sorts.

'In the final acquisition of our IP at the end, it was the patent portfolio that proved key'

Reprieve

Behind it all, though, NVMdurance had developed some significant engineering milestones. The false dawns listed above were only ever evident due to the truly appealing nature of the inventors’ discoveries.

And the most important thing Coyle, Sullivan and Ryan did was prioritise patenting. This is not a difficult procedure, but it certainly is expensive. What helped finance this process was the series of significant paid trials secured with industry leaders.

Lawyering in this space is competitive. Are your patents better than mine? Can my patents side-step yours etc?

Each country, each territory has local, excellent IP lawyers. However, once you decide you want to operate in an industry as specific as semiconductors, Coyle says Silicon Valley is the only place to look.

“We got advice early on that if we were going to get acquired, there would be huge regard to the patents we had. In the semiconductor industry, companies file patents a lot. They do it to put companies out of business. It was important we had offensive and defensive patents.

“We looked for, and got an IP lawyer with great credentials in memory. He took what was our badly drafted patent, and found ways to change it, enhance it, spawning a whole bunch of patents off the back of that.”

But what good are patents if no one will pay?

With no real urgency, Coyle and his team dealt with their shareholders and made for the door. Then, “out of the blue”, an offer came in late 2017.

“An IP legal firm that was representing someone – and we still don’t know who – made an attractive offer to buy the IP. Then we weren’t responding quick enough, and they wanted the deal done by the end of the calendar year, so they upped the price to make it even more attractive.

“In the final acquisition of our IP at the end, it was the patent portfolio that proved key. Patent filings were our second biggest spend after payroll. That was a big leap of faith. But as it turned out we would have had nothing if we hadn’t done that.

“We had an exit, of sorts.”

'But that’s the game…'

Learnings

Now, several months down the line, the inventors are back in the memory space and patenting, still as NVMdurance. The pivot finally happened.

Coyle has raised a deep-tech investment fund, alongside Burniece. At the very least, he’s armed with an awful lot more learnings.

“I’ve been in this game a while. I have had ups and downs. One of the main things I learned, though, was managing expectations. During one of the times we were raising money, one of our current investors was approached by a potential future one. I found out that, among the discussions, the new investors asked what it was like to work with us. Our investors said, ‘Pearse always tells you the bad news as early as possible’. I learned that that was a trait that investors, everyone, liked. Managing expectations was key.

“Our staff were always really keenly interested in general progress and funding. I remember telling them about the convincing verbal offer of VC funding meaning we could carry on. Then, having to then tell them that no, sorry we were wrong. That was very, very hard.

“But that’s the game…”