How AMD Boosted Hedge Accuracy, Minimized FX Volatility

“We run in a daily operating rate environment, so anything that changes our balance sheet is effective immediately,” Carbone says. Retrieving the trial balance, configuring the data in spreadsheets, and analyzing exposures took the team approximately two hours a day. Plus, once a month, they would calculate each hedge’s mark-to-market value and test its effectiveness, which took several more hours. These processes were so time-consuming that the team evaluated—and hedged—only the company’s biggest exposures, while they lacked insights into other risks that may have been material to the balance sheet.

Cash flow hedging was equally manual, although it was more manageable on a daily basis because the team placed cash flow hedges only once a month. “We would take historical operating expenses and work with FP&A [financial planning and analysis] to figure out growth rates, based on workforce planning, for different areas of the business,” Carbone says. “We would incorporate those growth expectations into our projections of exposures over the next year or two. Then we would initiate the appropriate hedges.”

All in all, AMD’s approach to FX risk management was inefficient, which created challenges in planning for the future and in integrating new entities into the corporate structure. The situation came to a head with a major acquisition that closed in 2022. The acquired business, Xilinx, utilized multiple functional currencies and employed a monthly operating rate. AMD opted to align the new, consolidated company on its legacy operational model, but its lean treasury group needed a more efficient way to manage financial risk.

When the acquisition occurred, AMD was midway through its deployment of the Kyriba FX Risk Management platform. The treasury team worked closely with Kyriba to establish more efficient FX risk management processes. They created automated feeds from SAP, Bloomberg, and their FXall trading platform into Kyriba.

“On the cash flow side, our policy allows us to hedge out to two years starting at a nominal hedge percentage and scaling it up as the cash flow date approaches,” Carbone explains. “We have flexibility to accelerate or decelerate the incremental hedging based on what spot, carry, and volatility are doing for a given currency. For example, in early 2025, when the U.S. dollar had strengthened to its highest level since 2022, we increased our hedges to policy limits to lock in that favorability for opex [operating expenses].”

For balance sheet hedging, “we will usually hedge exposures over a certain threshold, but we will take into consideration whether the carry cost is too high compared with the level of volatility,” he continues. “We will also see how a given currency is trending against the U.S. dollar. Based on that, we will determine whether we want to make any changes from a hedge percentage standpoint on a given currency. We usually stick to somewhere between 80 percent and 100 percent for most currencies, but we will true up those percentages each month.”

In both processes, once the team decides where it needs to place FX hedges, it enters them into the Kyriba system. They are routed to the treasurer for approval, then the team can push them directly into FXall for execution. “For cash flow hedging, we are not pegged to one date; we like to spread out the transactions to get a weighted average price over time,” Carbone says. “After a trade is executed, FXall pushes the information back into Kyriba, which recognizes it and tags it accordingly. That has significantly reduced our workload in terms of actually placing the hedges.”

Not only are currency hedges now streamlined, but having all the company’s FX exposure and hedging data in a single platform enabled the treasury team to transfer hedge accounting to AMD’s accounting department. “Treasury used to own hedge accounting, but since the new system is mostly automated, the accounting team was able to take over that work,” Carbone says. “They can pick up all the month’s hedges at month-end and run the mark-to-market analysis. Our new process meets all our control requirements and provides better visibility to all aspects of FX risk management, which makes it easier to have discussions with our auditors.”

All told, treasury has been able to reduce the multiple hours they previously spent every day on balance sheet hedging down to about 30 minutes. This has enabled them to shift their attention from day-to-day operational tasks toward more strategic FX hedging initiatives—one of which was deploying an efficient-frontier tool designed to help optimize the balance between FX risk and hedging costs.

“We’ve freed up time to focus on strategic aspects of FX,” Carbone says. “Our efficient-frontier tool enables us to put in our exposures and our target for value at risk [VaR], and then it will tell us what percentages we should hedge in each currency. We’ve used the tool to reduce our annual hedging costs by around $2 million, with minimal increase in VaR.”

The efficient-frontier perspective has also enabled the team to make better decisions about where FX risk management will be most effective. “For example, we used the efficient-frontier tool to look at the Taiwan dollar,” Carbone says. “It has a carry cost of about 4 percent, which is high for a currency whose volatility is around 8 percent. We looked at whether it was worth hedging that level of volatility for that embedded cost. Ultimately, we decided we shouldn’t be hedging the Taiwan dollar, and we were able to extract some cost savings by unwinding those hedges.”

The team has also started looking at new instruments—primarily options and collars—for their cash flow hedging. An FX dashboard that they built in Microsoft Power BI “reports our P&L impact for the quarter, both for the balance sheet and cash flows,” Carbone explains. “It also includes trade statistics like counterparties’ performance. We work with 10 different banks on FX hedges, and we use the dashboard to look at the cost differentials between each bank’s bids and the winning bids. Then we have an annual performance review with each bank so they can see how they’re performing in each currency. That’s been helpful in reducing our FX pricing.”

Even more important, the automation in the new system enables the team to look at a broader range of risks. “Now we can put our whole balance sheet in Kyriba to see all those exposures—some of which we weren’t previously capturing,” Carbone says. “If something changes with an exposure, we can dig into SAP to see what’s causing the variances. And because we can see the full exposure of every currency coming from the ERP system, we can make better decisions about hedging. We estimate that, as a result of all our process changes, we’ve reduced the overall P&L impact of FX by approximately 60 percent.”

One more benefit of the new system is that it streamlined AMD’s integration of Xilinx into its treasury processes. “We set one FX hedging policy for the whole company, which required a lot of changes for the acquired organization,” Carbone reports. “Most of their entities outside the United States were non–USD-functional. There were also differences in the ways we accounted for certain things, so we had to make sure all the journal entries would be produced correctly. Having a hedging system that automated a lot of those workflows made the integration of our currency risk management processes a lot easier.”

Overall, Carbone concludes, the transformation of AMD’s FX processes was intensive but well worth the effort. “Our new approach to FX risk management has made our hedging a lot more accurate and reduced the volatility of FX into our OI&E [other income and expense] and opex,” he says. “To other treasury teams that are stuck doing FX risk management in Excel, I would say that if you have a high enough hedging volume and low transparency to exposures from your ERP system, there is a huge benefit of finding a partner that will help you automate a lot of those processes.”