Leadership: Lisa Su and AMD

Unit 3, Leadership: Lisa Su and AMD.

Why does a company that once made excellent products end up in trouble? And how does the person who takes it over reorganize the capabilities that still exist into a path that actually works?

Here goes — enjoy:

Case Summary

AMD's early opportunity came from customers' need for a second source of supply. IBM didn't want to rely on a single trusted source for a critical processor, which gave AMD an important way into the PC industry. It could offer customers choice while working with the software ecosystem that already existed. [1]

But AMD's history also includes moments when it defined products itself. The Athlon, which began shipping in 1999, and later the products that reached 1 GHz, proved its engineering strength. The Opteron, launched in 2003, let customers upgrade to 64-bit computing while continuing to run their existing 32-bit software. Compatibility protected customers' past investments, and also opened the door to new technology. [2] [3] [4]

Opportunity came with bigger commitments. In 2006 AMD completed its acquisition of ATI, at an announced transaction value of about US$5.4 billion, taking out a US$2.5 billion term loan for the deal. Graphics capability came into the company, and so did pressure on cash and debt. In 2009 AMD spun off its wafer manufacturing business as GLOBALFOUNDRIES, beginning to change the way it carried the cost of manufacturing. [5] [6]

Meanwhile, its products ran into competitiveness problems. Contemporary benchmarks of the FX-8150 in 2011 showed that single-threaded performance dragged down the experience in some applications and games. Having more technical ideas doesn't automatically mean customers get better results. [7]

By the time Lisa Su took over as CEO in October 2014, the company already had some foundations worth keeping: the CPU team was being rebuilt, the semi-custom game-console business already had customers, and the manufacturing spin-off had long been completed. Keller had returned in 2012, and the use of AMD chips in Sony's and Microsoft's new generation of consoles had been announced in 2013. What she inherited was a set of capabilities and commitments that needed to be reconfigured. [8] [9] [10]

The quarterly snapshot around the time she took over: cash, cash equivalents and marketable securities of about US$938 million, and total debt of about US$2.2 billion. The traditional Computing and Graphics segment was under pressure, while the Enterprise, Embedded and Semi-Custom segment was profitable. Not all the debt was due immediately, and these figures can't be converted directly into "how many months it can survive"; what they suggest is that room for allocating resources was already limited. [11]

Freezing time at this point, I'd frame the choices in front of her as three questions:

1. With limited cash, which spending do you stop first, and which capabilities do you protect?
2. How do you set the order of investment among CPUs, GPUs and the existing customer businesses?
3. How do you get employees, customers and the board to jointly accept a plan that will take years to prove itself?

My Thinking

Technology that once led needs an organization that keeps delivering

AMD's early history shows that technical capability can change a company's position. Customers are willing to seriously consider a challenger because it offers concrete value — performance, compatibility, supply choice, or reasonable cost.

But a lead doesn't last. Whether the next-generation product comes out on time, whether the software is ready, whether suppliers can cooperate, whether customers are willing to adopt it again — all of this requires the organization to keep delivering.

My own understanding is that a technical breakthrough has to be caught by a whole set of delivery capabilities. Technology opens up possibilities; the organization makes sure those possibilities still hold the next time, and the time after that.

Cash buys time for long-term technology

If I put myself in 2014, I'd look first at cash flow, payment commitments and R&D milestones. A new processor's development cycle doesn't suddenly get shorter because there's a new CEO.

Stopping some work might bring expenses down. But if what gets stopped is exactly the key capability the next-generation product needs, the company just gets smaller without gaining any new reason to compete.

My choice would be to put the short-term cash plan and the long-term technology plan on the same table. For the businesses that can contribute resources today, assess how much time they can buy for tomorrow; for the technologies tomorrow's hopes rest on, require progress that can be checked today.

There's only so much time in the hourglass. The leader has to decide which things are worth letting it run out on slowly.

Focus needs a real stop list

CPUs, GPUs, embedded, game consoles — every direction comes with an opportunity story. The problem is that every opportunity takes up engineers, management attention and customer-support capacity.

If I were Lisa Su, I'd make high-performance computing the main direction, review the core R&D already under way, and maintain the valuable existing customer businesses. Projects off the main line would need stronger customer evidence, tighter investment limits, and clear stopping conditions.

Focus ultimately shows up in resource allocation. Who gets the best team, who gets continued budget, who has to stop — that's where strategy actually happens.

My own inclination is to pursue fewer directions and accept some opportunity cost. This choice carries risk too: the bet becomes more concentrated, so it needs staged validation all the more.

Acknowledging what your predecessors built is how you put resources in the right place

Looking back at a success story, it's easy to rearrange the whole history around one person. The change in manufacturing model, the return of key talent, the winning of important customers all seem to have happened only after the new leader arrived.

AMD's timeline reminds me that a successor's first job is to judge what's already in hand. What's truly valuable, what needs continued investment, and what — even though a lot of money has gone into it — should be stopped.

Accurately acknowledging what the organization has built helps keep the people who actually do the work. A CEO can bring new decisiveness, and also needs to respect the people and the work that make those decisions possible.

My understanding is that part of leadership lies in organizing what's been built up across different teams into a direction everyone is willing to share responsibility for.

Trust has to be rebuilt through round after round of delivery

After the strategy is announced, the truly fiddly work begins: performance misses targets, validation hits problems, customer requirements change, the supply chain needs coordinating.

I'd give every major product a clear owner, and have R&D, sales and operations face the same set of facts together. When problems come up, first get the real situation into the meeting, then discuss responsibility and remedies. The earlier bad news is reported, the more choices the team has.

My first three steps would be these: get a clear picture of the facts and the cash, concentrate resources, and then deliver on commitments one by one. The three move forward in parallel, and within the first 90 days there should be visible changes in plans, accountability and how things are validated.

The long-term direction deserves patience; specific projects need testing. Only by doing both well at once do employees have a reason to keep investing themselves, and customers a reason to give you another chance.

Takeaways for Arcus42

What this case gives me is a few reminders about how to build an organization.

First, be clear about the core capabilities worth building up over the long term. A service may bring in immediate revenue, but also look at whether it improves judgment, delivery quality and client trust. Before a new opportunity comes in, check how it relates to that capability.

Second, give experiments a budget, and an end condition. Leave room for new ideas, while agreeing on when to review them and what evidence is needed to continue. The team's attention needs to be allocated carefully too.

Third, protect the work that short-term reports can't yet fully reflect. Knowledge curation, quality review and developing people all take time. Protecting them also means explaining how they improve delivery, and continually checking on progress.

Fourth, turn client commitments into a shared working standard. Who's responsible, when it's due, how it's signed off — all of this needs to be clear enough. Getting problems to surface early is usually worth more than explaining them afterwards.

Final Thoughts

Before class, my vote goes to this line: leadership in a crisis means making limited resources support a path that can keep delivering.

---

A note on sources: this piece is based on publicly available historical material. The assigned version of The Legacy of AMD mentioned by the professor has not yet been verified; "my choice" in the text is pre-class analysis and does not represent the conclusions of the professor or the assigned materials. The financial snapshot comes from quarterly results for the period ended September 27, 2014, published on October 16 of that year. This piece discusses business management and does not constitute investment advice.

Source index:

[1] Computer History Museum, Intel 386 Business Strategy oral history, 2008.

[2] AMD, 1999 Annual Report, filed 2000.

[3] AMD, 2000 results announcement, filed 2001.

[4] AMD, 2002 10-K, filed 2003.

[5] AMD, announcement of completion of ATI acquisition, October 25, 2006.

[6] AMD, 8-K, March 5, 2009.

[7] Clive Webster, AMD FX-8150 Review, bit-tech, October 12, 2011.

[8] AMD, announcement of Jim Keller joining the CPU team, August 1, 2012.

[9] AMD, 2013 10-K, filed 2014.

[10] AMD, announcement of Lisa Su's appointment, October 8, 2014.

[11] AMD, Q3 2014 results announcement, October 16, 2014.