It’s hard to believe it’s already been a year since Intel’s favorite son Pat Gelsinger ousted CEO Bob Swan to take the company’s reigns.

To call Gelsinger’s strategy aggressive would be an understatement, and to his credit, the Intel of today is a very different animal compared to that of a year ago.

As a 30-year Intel veteran, engineer, and former CTO, Gelsinger was given license to make some bold moves, and he’s certainly wasted little time bending the company to his will.

How many CEOs do you know in their first month have dropped $20-plus billion to grow the company’s footprint at a time when its tech was rapidly falling behind and investor confidence was at an all-time low? And that was only the beginning.

So far, big and bold seems to be paying off. Intel is full of brilliant scientists and engineers, but the company has spent the better part of a decade resting on its laurels while rivals at AMD, Amazon, Ampere, and Marvell have steadily chipped away at Intel’s lead.

Investing in innovation, in my opinion, was exactly what the company needed.

To be perfectly clear, Intel still faces its stiffest competition in decades and Gelsinger really does have his work cut out for him.

Many, if not all, of the criticisms I leveled against the company nearly two years ago still hold true, and the challenges facing the chipmaker have only grown more numerous.

Building a foundry business to rival Taiwan Semiconductor Manufacturing Co. (TSMC) or Samsung Electronics isn’t going to be easy, let alone cheap, especially when both companies are outspending Intel by a considerable margin.

While I commend Gelsinger’s strategy, I am beginning to wonder how long he can keep it up. He is burning through capital to support Intel’s fledgling foundry business, but the company won’t see any profit from those investments until the middle of the decade at the earliest. Intel’s Arizona and Ohio foundry expansions won’t come online for another two or three years.

What’s more, he’s fronted the cash in anticipation of a $52 billion bailout from congress, which has, despite his pleas, so far failed to materialize.

It’s a big bet, but one that I’m convinced will ultimately pay off. I don’t see semiconductor demand easing any time soon, especially given the geo-political trend toward domestic manufacturing and supply chain security.

It’s going to be a while before investors, accustomed to massive gross margins, will see Intel rise to its former glory. And having watched confidence in Intel erode so rapidly in Swan’s final days, it will only take a few missteps to shake investors’ faith again.

That pressure appears to be getting to Gelsinger, who in an interview with Bloomberg Television last week, slammed Wall Street for doubting his vision.

It’s also important to remember that Intel is still falling behind its competition in both the data center and consumer spaces.

Intel’s ill-fated 7-nanometer manufacturing process is still more than a year out. And while you can’t directly compare one process node to the next, conventional wisdom holds that Intel’s 10-nanometer manufacturing process is roughly on par with TSMC’s 7-nanometer process, at least in terms of transistor density.

Based on this, we can expect the upcoming process to offer similar performance to AMD’s EPYC 4, which is slated for release this year.

Making matters worse, by the time Intel can bring it’s 7-nanometer process — since rebranded as Intel 4 — to market, TSMC will already be ramping production of a 3-nanometer process.

There is some light at the end of the tunnel for Intel though.

The company is aggressively pursuing a novel 2-nanometer — or 20 angstrom — manufacturing process that’s unlike anything the company has made before. It uses a transistor architecture similar to one developed by IBM and Samsung last year, called RibbonFET. The technology stacks transistor gates vertically to achieve higher compute densities than previously possible.

But like its foundry projects, the new process isn’t slated for release until 2024 at the earliest, and given Intel’s track record for meeting deadlines, I’m not holding my breath.

Intel still has a difficult road ahead of it, but it appears so far that Gelsinger is more than capable of tackling these challenges.

If Intel fails at this point, it will be for lack of faith, not for lack of vision.