Caffeine for Peak Performance: 2.7% Increase in Max-Power Can Make All the Difference | Plus: Timing Matters! "The Caffeine Buzz" Occurs 30 Min After Blood Levels Peak

Yes, caffeine "doping" may in fact allow you to show up at the office in time, even when you've overslept (only useful if your tiredness is not due to a caffeine-abuse induced lack of sleep, obviously).

It sounds unbelievable, but up to now most of the research into the effects of caffeine on single bouts of brief ( ≤30 s) maximal exercise, predominantly using 30-s sprint cycling tests, shows no effect:  Bell et al. (2001), Collomp et al. (1991), Glaister et al. (2012), ... the list goes on. None of these and a bunch of other studies found increases in sprint performance irrespective of the amount and mode of caffeine supplementation.

Until today, only Anselme, Collomp, Mercier, Ahmaidi, and Prefaut (1992) found a significant effect of caffeine on maximal anaerobic power output (Wmax), as derived from a series of maximal 6-s cycle sprint tests. Unfortunately, study by Anselme et al. (1992) has some limitations including: (1) the use of a mixed gender sample; (2) the use of a fixed (250 mg), rather than a body mass-relative caffeine dose; (3) the absence of serum caffeine analysis to confirm caffeine abstinence; and (4) the absence of a familiarisation trial.
The aim of Mark Glaister and his colleagues from the St. Mary's University in Twickenham, UK, was thus to "repeat the study by Anselme et al. (1992), addressing the aforementioned issues, in an attempt to provide a clear answer as to whether caffeine has an effect on sprint cycling performance" (Glaister. 2014).

Update: The Latest on Caffeine, Exercise, Fat & Weight Loss | more

It would be beside if I tried to keep you on the tenterhooks. From the headline of today's SuppVersity article you know after all that the experiment was a success. Glaister and his colleagues whose experimental protocol involved fourteen male Strength and Conditioning and Sport Science students, who were regularly active in strenuous physical activity instead of average coach potatoes (that's important, because the results will differ), was a success.

The scientists were able to show that caffeine will actually increase peak anaerobic power output in a series of 6-s cycle ergometer sprints, separated by 5-min passive recovery periods.

Figure 1: Statistical significant performance increases occur only at torques that allow the subjects to perform at their individual maximal aerobic power output (W_max) - torques that were not used in previous studies (Glaister. 2014)

As you can see in Figure 1 the differences which reached statistical significance only on the latter of the sprints were not earth-shatteringly large, but they were there and could very well make the difference between victory and defeat in any competitive athlete.
References:

• Anselme, F., et al. "Caffeine increases maximal anaerobic power and blood lactate concentration." European journal of applied physiology and occupational physiology 65.2 (1992): 188-191.
• Bell, Douglas G., I. R. A. Jacobs, and K. Ellerington. "Effect of caffeine and ephedrine ingestion on anaerobic exercise performance." Medicine and science in sports and exercise 33.8 (2001): 1399-1403.
• Collomp, K., et al. "Effects of caffeine ingestion on performance and anaerobic metabolism during the Wingate test." International journal of sports medicine 12.05 (1991): 439-443.
• Cox, Gregory R., et al. "Effect of different protocols of caffeine intake on metabolism and endurance performance." Journal of Applied Physiology 93.3 (2002): 990-999.
• Glaister, Mark, et al. "Caffeine and sprinting performance: dose responses and efficacy." The Journal of Strength & Conditioning Research 26.4 (2012): 1001-1005. 
• Glaister, Mark, et al. "Caffeine supplementation and peak anaerobic power output." European journal of sport science ahead-of-print (2014): 1-7.
• Kaplan, Gary B., et al. "Dose‐Dependent Pharmacokinetics and Psychomotor Effects of Caffeine in Humans." The Journal of Clinical Pharmacology 37.8 (1997): 693-703.